Oh ...
"nanobots would be able to attach to cells, tell if they were malignant, and if so kill them"
Milspec version --> "nanobots would be able to attach to cells and kill them"
American boffins say that they have created tiny robot spiders or "walkers", each one of which is a single complicated molecule fashioned out of DNA. They have managed to get the molecular nanorobospiders to follow a trail of DNA "breadcrumbs". DNA walker nanobot in action. Credit: Paul Michelotti The sticking-up bits are …
That's when KE Drexlers "Engines of Creation" was first published.
People are *still* concerned about gray goo. Did you *really* think you were the first to coin that phrase?
Self replication is a pretty tough problem. However in chemistry and biology its *fairly* easy to make *lots* of copies of a component molecule (6x 10^23 parts of *anything* but molecules or bacteria is a *lot* of something. In biology and chemistry its a common number of manufactured "parts"). Making them self assemble (or creating an environment which *encourages* them to self-assemble) is *much* easier.
BTW DNA is being use for several reasons a)Techniques exist to mfg a chain at will, roughly at £0.4/base pair then replicate it in very large numbers (PCR).2)DNA is *more* than a kind of molecular mag tape. Chunks of it can control interpretation of following sections. The ability to control the actions of the tool molecules which bond to it (which it can be exposed to in different sequences, add another degree of freedom to the process) make it akin to a physical Turning machine. c)The range of tool molecules is *extensive*.
As for developments taking "100 years" I think 100 months is more likely.
Mine's the one with Martyn Amos's "Genesis Machines" in the pocket.